Wildlife fences have been proven to be effective in reducing deer-vehicle collisions (DVC), if properly designed, installed and of sufficient length. Fences, however, cannot be installed everywhere due to, e.g., the costs of installation and maintenance or esthetical reasons. For this reason several less costly and less conspicuous mitigation measures have been developed. One of these are wildlife reflectors, which can be installed in road verges. They are supposed to reflect the lights of passing cars into the surrounding habitat, aiming to stop deer from crossing the road when there is traffic. Currently a variety of wildlife reflectors are on the market. One of the more recent types are blue wildlife reflectors. In the Netherlands, these blue reflectors were quickly adopted by road managers and hunters, as they perceived them to be effective. Within a few years hundreds of roads have been equipped with these blue wildlife reflectors. Yet, no scientific study has been carried out to support the claims that the blue reflectors reduce DVC. Our objective is to test whether blue wildlife reflectors reduce DVC, and if so, to what extent. We collected roadkill data of roe deer (Capreolus capreolus) on 37 roads (total road length: 26.8 km), both before and after the installation of the blue wildlife reflectors. As the year of installation differed between roads, the number of years before mitigation varied between 5 and 8 years, and the number of years after mitigation varied between 2 and 5 years. Therefore, we analysed two scenarios: (1) 5 years before and 2 years after (n=37 roads), and (2) 5 years before and 5 years after (n=18 roads). We used the two-sample Poisson-test to assess whether a statistically significant (p<0.05 significance level) change in DVC before and after mitigation occurs. Our null hypothesis was that the average number of DVC before and after mitigation does not differ. For both scenarios we could not reject the null hypothesis. In scenario 1 average yearly roadkill was, respectively, 24.4 and 24.5 before and after mitigation. In scenario 2 average yearly roadkill was, respectively, 24.4 and 24.2 before and after mitigation. Population numbers in the areas surrounding the studied roads declined with 9% (scenario 1) and 17% (scenario 2) after mitigation. If we correct for these trends in population numbers, assuming a linear relation between population numbers and roadkill, we found a trend of 10% (scenario 1) and 19% (scenario 2) increase of roadkill after the reflectors had been installed. Our study emphasizes the need for evaluations of road mitigation measures prior to large-scale implementation. It shows that signals from practitioners about mitigation effectiveness can be biased and therefore should be treated with care. Our findings reflect the outcome of recent studies carried out elsewhere in Europe. Hence, no green light for blue wildlife reflectors yet.